Multi-scene accurate sensing method and device for new energy high-permeability power system

Abstract

The invention discloses a multi-scene accurate sensing method and device for a new energy high-permeability power system. The method comprises the following steps: acquiring power data of each side; adjusting the output of the power supply side based on the power data and the carbon budget constraint on the two sides of the source network; if it is judged that low-carbon demand response is triggered based on the power data on the two sides of the network load, the power utilization behavior on the load side is adjusted based on a demand response excitation mechanism; determining a cooperative control strategy based on the power data at the two sides of the source storage; the energy storage side is optimized and controlled based on the power data of the power grid side, the response index is assessed based on the power data of the energy storage side, and the energy storage side is controlled to participate in the auxiliary service of the power grid side based on the assessment result; and monthly regulation, day-ahead regulation and intra-day regulation are determined based on the power data of the source network load storage sides. Therefore, accurate sensing of multiple scenes of the power system can be realized, optimized operation of the power system can be guided, and safe, efficient and low-carbon operation of the power system is supported.

Description

Technical Field

[0001] This application relates to the field of power system technology, and in particular to a method and device for accurate sensing in multiple scenarios of power systems with high penetration of new energy sources. Background Technology

[0002] As the power system transforms into a new type of power system with a high proportion of new energy access and complementary multiple energy sources, traditional methods are difficult to meet the needs of accurate sensing in various scenarios of power source, grid, load and storage. Therefore, there is an urgent need for an accurate sensing method that can achieve full-domain and coordinated operation to solve the problem of accurate sensing in multiple scenarios of the power system. Summary of the Invention

[0003] This application provides a method and device for precise sensing in multiple scenarios of a power system with high penetration of new energy sources. It can realize panoramic monitoring of the power system, achieve precise sensing in multiple scenarios of the power system, guide the optimized operation of the power system, and support the safe, efficient and low-carbon operation of the power system.

[0004] In a first aspect, embodiments of this application provide a method for accurate sensing in multiple scenarios of a power system with high penetration of new energy sources, including: Acquire power data from the power source side, grid side, load side, and energy storage side of the power system; The power output of the power source is adjusted based on the power data from the power source side, carbon budget constraints, and power data from the power grid side. If a low-carbon demand response is triggered based on the power data from the power grid side and the power data from the load side, the electricity consumption behavior of the load side is adjusted based on the demand response incentive mechanism. Based on the power data from the power source and the power data from the energy storage side, a coordinated control strategy for the power source and the energy storage side is determined. The energy storage side is optimized and controlled based on the power data of the grid side, and the response indicators of the energy storage side are evaluated based on the power data of the energy storage side. Based on the evaluation results, the energy storage side is controlled to participate in the auxiliary services of the grid side. The monthly, day-ahead, and intraday controls for the power supply and load sides are determined based on the power data from the power source side, the carbon budget constraint, and the power data from the load side.

[0005] Secondly, embodiments of this application provide a multi-scenario precision sensing device for power systems with high penetration of new energy sources, including: The acquisition module is used to acquire power data from the power source side, grid side, load side, and energy storage side of the power system. The source-load linkage module is used to adjust the output of the power source based on the power data, carbon budget constraints and power data of the power grid. The grid-load linkage module is used to adjust the electricity consumption behavior of the load side based on the power data of the grid side and the power data of the load side if it is determined that a low-carbon demand response is triggered. The source-storage linkage module is used to determine the collaborative control strategy of the power source side and the energy storage side based on the power data of the power source side and the power data of the energy storage side. The grid-storage linkage module is used to optimize and control the energy storage side based on the power data of the grid side, to evaluate the response indicators of the energy storage side based on the power data of the energy storage side, and to control the energy storage side to participate in the auxiliary services of the grid side based on the evaluation results. The source-load linkage module is used to determine the monthly, day-ahead, and intraday regulation of the power source and the load side based on the power data of the power source side, the carbon budget constraint, and the power data of the load side.

[0006] Thirdly, embodiments of this application provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the method provided in embodiments of this application.

[0007] The technical solution provided in this application can achieve panoramic monitoring of the power system by acquiring power data from the power source side, grid side, load side, and energy storage side. Based on the acquired power data, it can realize source-grid linkage, grid-load linkage, source-storage linkage, grid-storage linkage, and source-load linkage, enabling accurate perception of multiple scenarios of the power system, optimizing the operation of the power system, and supporting the safe, efficient, and low-carbon operation of the power system. By using a cross-attention mechanism to fuse multimodal data of the power system, and incorporating cross-side correlation patterns, it can accurately realize fault diagnosis, new energy power prediction, and low-carbon regulation, and solve the problems of complex data processing and accurate perception in multiple scenarios of the power system. Attached Figure Description

[0008] Figure 1 A flowchart illustrating a method for precise sensing in multiple scenarios within a power system with high penetration of new energy sources, provided in this application embodiment; Figure 2 A block diagram of the power system's overall perception capabilities; Figure 3 Framework diagram for a multi-scenario accurate perception system; Figure 4 This is a schematic diagram of the interconnection between different sides of the power system; Figure 5A comprehensive framework diagram of a power system-wide, layered, and interconnected source-grid-load-storage precision sensing system; Figure 6 This is a schematic diagram of the BGRU structure; Figure 7 A structural block diagram of a multi-scenario precision sensing device for a power system with high penetration of new energy provided in this application embodiment; Figure 8 This is a schematic diagram of an electronic device structure provided in an embodiment of this application. Detailed Implementation

[0009] The present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0010] Figure 1 This application provides a method for precise sensing in multiple scenarios of a new energy high-penetration power system. The method can be executed by a device for precise sensing and linkage in multiple scenarios of a new energy high-penetration power system. The device can be implemented by software and / or hardware. The device can be configured in a corresponding system. The system adopts a hierarchical setting. The device can be configured in an electronic device.

[0011] like Figure 1 As shown, the method provided in this application embodiment may include the following steps: S110: Acquire power data from the power source side, grid side, load side, and energy storage side of the power system.

[0012] In this embodiment, the power system can be sensed through a multi-scenario precision sensing system. For a comprehensive sensing scenario, please refer to... Figure 2 For power source-side sensing, power data can be collected using sensing devices on the power source side. Specifically, this can be done using sensing devices in photovoltaic power plants, wind farms, and hydropower stations to collect data such as carbon dioxide levels, spatiotemporal parameters, environmental parameters, and physical parameters. In photovoltaic power plants, power generation parameters can be obtained using terminals, meteorological parameters such as light intensity, temperature, and wind speed can be collected using installed instruments, and equipment appearance can be inspected using cameras. In wind farms, operating parameters such as wind turbine vibration, temperature, and speed can be obtained using deployed wind turbine condition monitoring systems, and monitoring can be implemented using equipment. In hydropower stations, comprehensive sensing of various operating parameters and equipment status on the power source side can be achieved through devices such as water level gauges, flow meters, and unit condition monitoring sensors.

[0013] In this embodiment, for grid-side sensing, power data can be collected based on grid-side sensing devices. Specifically, electrical parameters and operating environment of the transmission line can be monitored in real time based on devices installed on transmission line towers. Within the substation, comprehensive sensing of substation equipment and operating environment can be achieved based on deployed automated protection devices, intelligent sensors (monitoring equipment temperature, pressure, partial discharge, and other status parameters), high-definition cameras, etc.

[0014] In this embodiment, for load-side sensing, power data can be collected based on load-side sensing devices and systems. Specifically, load power, power factor, voltage and current waveforms, and production process-related parameters can be collected in real time based on corresponding devices and systems. For residential loads, smart meters can be used to automatically collect electricity consumption information, and the use of energy management systems can be promoted to monitor the electricity status and energy consumption of household appliances.

[0015] In this embodiment, for energy storage-side sensing, power data can be collected based on sensing devices on the energy storage side. Specifically, parameters such as battery cell voltage, temperature, and charging / discharging current can be monitored using relevant devices to achieve comprehensive and multi-level accurate sensing of the energy storage system.

[0016] In this embodiment, the multi-scene precise perception system can adopt a layered architecture, which can be referred to as... Figure 3The layered architecture can include the end (perception layer), edge (edge ​​layer), network (transmission layer), cloud (platform layer), and intelligence (intelligent decision-making layer). The perception layer can be composed of various sensing devices to achieve multi-dimensional sensing of carbon dioxide levels, spatiotemporal quantities, environmental quantities, physical quantities, and state quantities, covering the entire process from source to network, load to storage. It transmits data to the edge layer via wired or wireless communication, providing raw data support for the upper layers. The edge layer can be based on a single edge device deployed near the sensing device, performing preliminary processing and analysis on the massive amounts of data collected. This includes filtering, aggregation, and feature extraction, reducing data transmission volume, and enabling on-site decision-making and execution of simple control tasks. The network transmission layer can construct diverse communication networks to establish data channels between the end, edge, and cloud, ensuring reliable and efficient transmission of sensing data and supporting end-to-end information interaction. The platform layer can receive pre-processed data from the edge layer and further store, manage, and deeply analyze the data. Based on advanced technologies such as dynamic power carbon emission factor calculation and diversified objective game theory, it performs global resource management, analysis, and decision-making, serving as the "data brain" of the overall architecture. It enables massive data storage and complex model computation, providing intelligent analysis and decision support for upper-layer applications. Specifically, the intelligent decision-making layer, supported by the platform's data, models, and algorithms, and driven by artificial intelligence algorithms, achieves functions such as panoramic monitoring of the power system, coordinated control of multiple power sources, prevention and control of large-scale power outage risks, and low-carbon intelligent management and control of new power systems. This guides the optimized operation of the power grid and supports the safe, efficient, and low-carbon operation of the power system.

[0017] S120: Adjust the output of the power source based on the power data of the power source side, the carbon budget constraint, and the power data of the grid side.

[0018] In this embodiment, regarding source-grid coordination, based on carbon budget constraints, a dynamic linkage mechanism is established between power generation output and grid-side scheduling by monitoring power supply output and grid-side line power flow and voltage information. When renewable energy output suddenly increases or decreases, the output of other power sources can be adjusted in a timely manner according to the real-time operating status of the grid, and the configuration of reactive power compensation devices can be optimized to ensure grid-side voltage stability and power balance. For example, during the peak output period of photovoltaic power plants at noon, if the grid-side load is relatively low, hydropower plants can be coordinated to reduce output in advance, while grid-side energy storage power stations can be arranged to charge, achieving coordinated operation between the source and grid.

[0019] Optionally, adjusting the power output of the power source based on the power data of the power source side, carbon budget constraints, and the power data of the power grid side includes: determining the output of new energy sources, line power flow redundancy, and carbon budget surplus based on the power data of the power source side and the power data of the power grid side; when the change in the output of new energy sources from the previous time point is greater than a preset amplitude threshold, and the line power flow redundancy is less than a preset redundancy threshold, scheduling the reactive power compensation device of the power source side, adjusting the output of the hydropower station of the power source side, and controlling the energy storage side to charge; if the output of new energy sources is lower than a preset output threshold and the carbon budget surplus is less than a preset target threshold, controlling the low-carbon operation of the thermal power units of the power source side and optimizing the line power flow transmission path of the power grid side.

[0020] Specifically, for source-grid linkage, data inputs can include: collecting power generation data from the power source side, line power flow data from the grid side, and voltage deviation data from substations, and calculating carbon surplus and line power flow redundancy. Linkage triggering conditions can be: the change in renewable energy output from the previous time point exceeding a preset threshold, line power flow redundancy falling below a preset threshold, and carbon budget surplus falling below a preset target threshold. The linkage execution process can be referenced as follows: (1) The edge layer can share power data from the power source side and the grid side to the platform layer, and the platform layer calculates the data correlation through the "carbon-power flow coupling model"; (2) Platform layer output differentiated control strategy: When the change in the output of new energy is greater than the preset amplitude threshold and the line power flow redundancy is less than the preset redundancy threshold, the reactive power compensation device on the power supply side is scheduled, the output of the hydropower station on the power supply side is adjusted, and the energy storage side is controlled to charge. If the output of new energy is lower than the preset output threshold and the carbon budget surplus is less than the preset target threshold, the thermal power unit on the power supply side is controlled to operate in a low-carbon manner and the line power flow transmission path on the grid side is optimized. (3) After receiving the control strategy, the edge layer sends it to the corresponding equipment (hydropower station speed governor, energy storage converter, reactive power compensation device). After execution, the equipment status is fed back to the intelligent decision layer. When the deviation exceeds the limit, the platform layer re-optimizes the strategy.

[0021] S130: If a low-carbon demand response is triggered based on the power data of the power grid side and the power data of the load side, the electricity consumption behavior of the load side is adjusted based on the demand response incentive mechanism.

[0022] In this embodiment, for grid-load linkage, a grid-load interaction mechanism can be constructed based on real-time operating data and carbon factor of the grid side and electricity consumption information of the load side. Through a demand response incentive mechanism, industrial load users are guided to adjust their production plans, reducing the electricity load of non-critical production equipment during peak grid load periods or increasing the electricity load during off-peak periods. At the same time, an energy management system on the load side is used to intelligently control the operating status of household appliances according to electricity price signals and real-time load conditions, thereby achieving peak shaving and valley filling and improving the economy and reliability of grid operation.

[0023] Optionally, if a low-carbon demand response is triggered based on the power data from the power grid and the power data from the load side, the electricity consumption behavior of the load side is adjusted based on the demand response incentive mechanism, including: determining the power grid load rate and the proportion of low-carbon power output based on the power data from the power grid and the power data from the load side; if the power grid load rate or the output of low-carbon power meets preset conditions, a low-carbon demand response is triggered, pushing peak load reduction instructions and off-peak production transfer suggestions to industrial users, pushing low-carbon electricity consumption periods and off-peak charging incentives to residential users, and increasing the compensation standard for residential users if the load adjustment amount on the power supply side does not meet the preset adjustment amount, and sending demand response benefit bills to residential users, and calculating the response rate of residential users, and classifying residential users based on the response rate of residential users.

[0024] Specifically, for grid-load linkage, data inputs can include: collecting grid-side load factor and time-of-use pricing signals from the sensing layer, industrial user electricity elasticity coefficients, residential user low-carbon electricity preference tags, and power output from the load side. Linkage triggering conditions can include: triggering when the grid-side load factor exceeds the peak threshold or falls below the valley threshold; triggering low-carbon demand response when the proportion of low-carbon power output exceeds the proportion threshold. The linkage execution process can include: (1) The edge layer is classified into "industrial adjustable load and residential flexible load": push peak load reduction instructions and off-peak production transfer suggestions to industrial users, and push low-carbon electricity consumption time reminders and off-peak charging incentives to residential users; (2) The platform layer can use the "load response monitoring model" to count the load adjustment amount. If the load adjustment amount is not reached, the compensation standard for residential users can be increased. For example, the preset adjustment amount can be 30. If the counted load adjustment amount is 20, the compensation standard for residential users can be increased so that residential users reduce their electricity consumption and the load adjustment amount increases to 30.

[0025] (3) The intelligent decision-making layer can push demand response benefit bills to residents monthly, and at the same time feed back the residents' response rate to the perception layer to optimize the next round of resident user classification. Among them, the benefit bill can be a benefit bill that compensates residents to make residents benefit, and the residents' response rate can be the residents' response rate to reminders or incentives for off-peak charging periods. Residents can be classified based on the response rate, and corresponding strategies can be executed based on the classification.

[0026] S140: Determine the collaborative control strategy for the power supply side and the energy storage side based on the power data, carbon budget constraints, and power data of the power supply side.

[0027] In this embodiment, a source-storage coordinated control strategy is formulated based on the carbon budget constraints, output characteristic data, and energy storage system status data of the power source. If the output of new energy sources fluctuates significantly, the energy storage side responds quickly, smoothing out power output fluctuations through charge and discharge control, thereby improving the stability and predictability of power output. For example, if a wind farm experiences a sharp drop in output due to a sudden change in wind speed, the energy storage system immediately discharges to supplement the wind power output shortfall, ensuring stable power output to the grid. During periods of no sunlight at night when the photovoltaic power station is not illuminated, the energy storage side can discharge according to grid demand and energy storage status, extending the effective power supply time of new energy sources.

[0028] Optionally, a coordinated control strategy for the power supply side and the energy storage side is determined based on the power data of the power supply side and the power data of the energy storage side, including: controlling the charging and discharging of the energy storage side if the predicted value of the new energy output fluctuation is greater than the predicted threshold; controlling the charging and discharging depth of the energy storage side if the health of the energy storage side is less than a preset health threshold; switching the backup energy storage unit of the energy storage side if it is determined from the power data of the energy storage side that the energy storage side has overcurrent or overtemperature; and optimizing the charging and discharging power allocation of the energy storage side if the fluctuation smoothing rate of the linkage between the power supply side and the energy storage side does not reach the preset standard.

[0029] Specifically, for source-storage linkage, data inputs may include: collecting predicted power output fluctuations and electricity generation carbon costs from the sensing layer, and battery health, charge / discharge carbon costs, and remaining capacity from the energy storage side. Linkage triggering conditions may include: triggering source-storage linkage when the predicted power output fluctuations of new energy sources exceed a preset threshold, or when the charge / discharge carbon cost of energy storage is lower than the difference between the charge / discharge carbon cost and the electricity generation carbon cost. The linkage execution process may specifically include: (1) The platform layer formulates a strategy through the “output fluctuation-energy storage response model”: when the output of new energy suddenly drops (the decrease in the output of new energy is greater than the preset threshold), the energy storage side is controlled to discharge rapidly; when the output of new energy suddenly increases (the increase in the output of new energy is greater than the preset threshold), the energy storage side is controlled to charge, and low-carbon cost power sources are given priority; when the battery health of the energy storage side is lower than the preset health threshold, the depth of charge and discharge is limited. (2) The edge layer can trigger protection commands and switch to backup energy storage units when overcurrent or overtemperature occurs based on the collected energy storage operation data; (3) The intelligent decision-making layer calculates the fluctuation rate of source-storage linkage every day and optimizes the charging and discharging power allocation when the preset standard is not reached.

[0030] S150: Optimize and control the energy storage device based on the power data of the grid side, evaluate the response indicators of the energy storage device based on the power data of the energy storage device, and control the energy storage device to participate in the auxiliary services of the grid side based on the evaluation results.

[0031] In this embodiment, for grid-storage linkage, the configuration and control of the energy storage side in the grid are optimized by combining the load characteristics and operating status data of the grid side. By configuring energy storage systems at key nodes of the grid, functions such as dynamic voltage support, power flow optimization, and emergency backup for the grid side are achieved. When a fault or abnormal voltage fluctuation occurs on the grid side, the energy storage side is controlled to quickly start operation, providing reactive power support and active power regulation to ensure the transient stability and power supply continuity of the grid. At the same time, the energy storage side can formulate reasonable charging and discharging plans based on the grid's time-of-use pricing and load peak-valley characteristics, participate in the grid-side peak-shaving and frequency regulation ancillary services, and improve the grid's operating efficiency and economy.

[0032] Optionally, the energy storage side is optimized and controlled based on the power data of the power grid side, including: if it is determined that the line power flow value of the power grid side is greater than the rated capacity ratio, controlling the energy storage side to charge; if the voltage deviation of the substation is greater than the qualified threshold, controlling the energy storage side to provide reactive power support; and if a fault occurs on the power supply side, controlling the energy storage side to discharge; and optimizing the deployment location and reserve capacity of the energy storage side based on the line power flow distribution of the power grid side.

[0033] Specifically, for grid-storage linkage, data inputs can include: data collected from the sensing layer on the grid side such as line power flow values, substation voltage deviation values, and fault signals; and on the energy storage side, data on charging and discharging response speed and reserve capacity. Linkage is triggered when the line power flow value exceeds the rated capacity percentage, the voltage deviation exceeds the acceptable threshold, or a fault occurs on the grid side. The linkage execution process can be as follows: (1) The platform layer can combine the “power flow-voltage-energy storage state model” to output strategies: when the line power flow value is greater than the rated capacity ratio, the energy storage side is controlled to charge; when the voltage deviation of the substation is greater than the qualified threshold, the energy storage side is controlled to provide reactive power support; when a fault occurs on the grid side, the energy storage side is controlled to discharge quickly to temporarily adjust the power flow and switch to the backup line. (2) The intelligent decision-making layer optimizes the deployment location and reserve capacity of the energy storage side based on the power flow distribution on the grid side every month; (3) The response indicators such as energy storage response time and power deviation are assessed, and energy storage that meets the standards is given priority to participate in grid-side ancillary services and receive benefits. Among them, grid-side ancillary services may include peak shaving and frequency regulation services.

[0034] S160: Determine the monthly, day-ahead, and intraday controls for the power supply side and the load side based on the power data from the power supply side, the carbon budget constraint, and the power data from the load side.

[0035] In this embodiment, regarding source-load linkage, under the constraints of carbon budget and guided by the electricity carbon factor, a monthly-day-day-intraday low-carbon regulation can be constructed. By coordinating the flexible interaction between the power supply side (generation side) and the load side (consumption side), the operation of the power system is optimized, the capacity for renewable energy absorption is improved, and the stability and economy of the power grid are enhanced. Through aggregation forms such as virtual power plants and smart microgrids, load fluctuations are effectively smoothed, peak-valley differences are reduced, and grid congestion is alleviated; a "balanced interaction" mechanism is constructed to enhance the system's regulation capacity and promote the efficient utilization of intermittent renewable energy sources such as wind power and photovoltaics; the grid's supply and demand balance capacity is enhanced, power supply reliability is significantly improved, and overall electricity costs are reduced through market-based means.

[0036] Optionally, determining monthly, day-ahead, and intraday regulation of the power supply side and the load side based on the power data of the power supply side, the carbon budget constraint, and the power data of the load side includes: determining the monthly carbon budget consumption rate based on the power data of the power supply side and the power data of the load side; if the monthly carbon budget consumption rate exceeds a preset planned value, determining the monthly low-carbon power output target based on the monthly carbon budget and the power supply side's carbon factor; controlling the monthly load capacity that can participate in peak shaving to match the monthly low-carbon power output target; and integrating industrial load and residential load into a negative load based on the adjustable load potential of the load side. The system aggregates loads and sends monthly control benchmarks to the power supply side and the load side; it determines the day-ahead coordination plan for the power supply side and the load side based on the next day's new energy output forecast and changes in the power exploration factor; it pushes the next day's adjustable load reservation instructions to the industrial load and pushes the next day's low-carbon electricity consumption period to the residential load; it adjusts the power supply side's output plan based on the load that can participate in regulation the next day; if the increase in new energy output leads to insufficient local load, it sends electricity consumption incentives to the load aggregate to guide the consumption of excess new energy output; and if grid-side congestion occurs, it controls the microgrid in the congested area to switch to islanded operation mode to prioritize the consumption of local new energy output.

[0037] Specifically, for source-load linkage, data inputs can include: power generation, real-time carbon factor, and daily / monthly carbon budget surplus from the power source side (wind power, solar power, thermal power, etc.) collected from the sensing layer; and power consumption and adjustable load potential from the load side (industrial and residential). The edge layer can calculate the renewable energy consumption gap and carbon budget consumption rate. Linkage can be triggered by meeting any of the following conditions: the monthly carbon budget consumption rate exceeds the preset planned value; the day-ahead renewable energy consumption gap requires guidance for load shifting; or intraday renewable energy output fluctuations lead to supply-demand imbalance. The specific linkage execution process can be as follows: Regarding monthly adjustments: (1) The platform layer can break down the monthly carbon budget according to the annual carbon target and calculate the monthly low-carbon power output target based on the carbon factor of each power source. (2) The edge layer can statistically analyze the adjustable load potential of each enterprise based on the sensing equipment on the industrial user side, and integrate the dispersed industrial load and residential load into a load aggregate through the virtual power plant aggregation platform to determine the monthly load capacity that can participate in peak shaving, so as to ensure that it matches the low-carbon output target on the power supply side. (3) The intelligent decision-making layer can issue monthly control benchmarks to the power supply side and the load side respectively: the power supply side specifies the maximum monthly carbon emissions of thermal power, and the load side specifies the monthly electricity consumption time limit for high-carbon processes.

[0038] Regarding the recent regulatory measures: (1) The platform layer can formulate a day-ahead coordination plan for the power supply side and the load side based on the next day's new energy output forecast and the dynamic changes of the carbon factor, and clarify the target for the proportion of low-carbon power output in each period of the next day; (2) The edge layer can push the next day's adjustable load reservation instruction to industrial users and the next day's low-carbon electricity consumption period to residential users through the management system; (3) The load aggregate feedback the amount of load that can be adjusted the next day. The platform layer can adjust the power output plan on the power supply side according to the load amount to eliminate the risk of supply and demand imbalance in advance.

[0039] Regarding intraday adjustments: (1) The perception layer can collect daily new energy output fluctuations and grid congestion signals in real time, and the edge layer will synchronize the collected data to the platform layer; (2) The platform layer can activate the balance interaction mechanism in real time by using the carbon factor: if the output of new energy sources suddenly increases and the local load is insufficient, the virtual power plant will send real-time electricity consumption incentives to the aggregated load body to guide the load side to start absorbing excess low-carbon output in time; if grid blockage occurs, the smart microgrid in the blockage area will switch to island operation mode to prioritize the absorption of local new energy output. (3) The edge layer can monitor the control effect in real time. If the new energy consumption rate does not reach the target, the load can be further guided to participate through market means to ensure the supply and demand balance between the power supply side and the load side and the implementation of the low-carbon target within the day.

[0040] The inter-system linkage framework in the power system can be referenced. Figure 4 The overall framework of a precise sensing system integrating the entire power system, its layers, and its interconnections can be referenced. Figure 5 .

[0041] The technical solution provided in this application can achieve panoramic monitoring of the power system by acquiring power data from the power source side, grid side, load side, and energy storage side. Based on the acquired power data, it can realize source-grid linkage, grid-load linkage, source-storage linkage, grid-storage linkage, and source-load linkage, enabling accurate perception of multiple scenarios of the power system, optimizing the operation of the power system, and supporting the safe, efficient, and low-carbon operation of the power system.

[0042] Based on the above embodiments, optionally, the power data can be multimodal data, including text modal data, frequency domain modal data, image modal data, and numerical modal data. For multimodal data, effective information can be extracted from the multimodal data, and the correlation between information in different modalities can be learned, thereby obtaining a unified representation of the multimodality.

[0043] The method provided in this application embodiment may further include: extracting features from the multimodal data and performing feature processing to obtain corresponding text modal features, frequency domain modal features, image modal features, and numerical modal features; for any one of the power supply side, grid side, load side, and energy storage side, determining the same-modal cross-side interaction features based on the text modal features, frequency domain modal features, image modal features, and numerical modal features using a cross-attention mechanism, and determining the corresponding single-modal fusion features based on the text modal features, frequency domain modal features, image modal features, numerical modal features, and the corresponding same-modal cross-side interaction features, and determining the global fusion features based on the single-modal fusion features and corresponding weights; splicing the global fusion features corresponding to the power supply side, grid side, load side, and energy storage side to obtain spliced ​​features, and performing fault detection, new energy power prediction, or low-carbon regulation based on the spliced ​​features.

[0044] Specifically, a bidirectional gated neural network (BGRU) can be used to extract features from multimodal data. The structure of BGRU can be found in [reference needed]. Figure 6 The gated recurrent unit (GRU) is used as the basic unit for extracting temporal features from multimodal data. Two GRU networks with forward and backward propagation are superimposed to perform bidirectional feature extraction.

[0045] The forward calculation formula is as follows: ; ; ; ; The reverse calculation formula is as follows: ; ; ; ; in, , The outputs of the update gate and reset gate at time t; It is the hidden state at time t; and These are the states at time t-1 and time t, respectively; , and It is the weight matrix connected to the input vector xt; , and It is connected to the state vector of the previous unit. The weight matrix; , and σ is the bias vector; σ is the activation function. Finally, the forward calculation yields... and the result of reverse calculation The results are summed up to obtain the final output of BGRU.

[0046] In this embodiment, text modal data X V Frequency domain modal data X F Image modal data X L Numerical modal data are X N， After extracting features for each modality, the feature vectors of all modalities are unified by feature dimension padding to obtain a standardized feature set. ;in, These are text modal features, frequency domain image modal features, image modal features, and numerical modal features. Using standardized multimodal features as input, a cross-attention mechanism is introduced to mine the correlations between source, network, load, and storage data of the same and cross-modal aspects, achieving bidirectional data empowerment.

[0047] Text modality features For example, extract the V from the power source side respectively. 源 , grid side V 网 , load side V 荷 Energy storage side V 储 Feature vector, V 源 As query vectors Q and V 网 V 荷 V 储 The column-wise concatenation is used as the key vector K and the row-wise concatenation is used as the value vector V. Substituting these values ​​into the cross-attention mechanism formula: ； in, Using this as the feature dimension, the cross-side interaction feature V of the power-side text modality feature is obtained through this calculation. 源-int Similarly, calculate V 网-int V 荷-int V 储-int Generate corresponding cross-side interaction features.

[0048] In this embodiment, the original feature information is retained, and an adaptive weight α (α∈[0,1], calculated in real time by the feature contribution evaluation model) can be introduced to obtain the modality fusion features. Taking the power supply side text features as an example: ； in, The single-mode fusion feature on the power supply side is used, and all mode features on other sides are fused according to the above logic.

[0049] In this embodiment, cross-attention is used to calculate the fusion feature weights for each modality (Q for a certain modality, and K and V for the others), and the weighted sum is used to obtain the one-sided global fusion feature, such as the power supply side: ； in, This represents the global fusion feature on the power supply side. 、 、 、 These are the power supply-side text modal fusion features, frequency domain modal fusion features, image modal fusion features, and numerical modal fusion features. Similarly, the power grid-side features are calculated. Load side Energy storage side .

[0050] The four-sided global fusion features are spliced ​​together in the order of "source-grid-load-storage" to form a unified global representation matrix of multi-scenario, multi-modal, source-grid-load-storage. This matrix retains key information from each mode while integrating cross-modal correlation patterns, providing direct data support for higher-level tasks such as fault diagnosis, renewable energy power prediction, and low-carbon regulation. Specifically, fault diagnosis, renewable energy power prediction, and low-carbon regulation can be performed using models based on the stitched features.

[0051] As the power system transforms into a new type of power system with a high proportion of renewable energy integration and multi-energy complementarity, the data scale of each link in the power system—source, grid, load, and storage—is growing increasingly massive, and the data types are becoming more diverse, including multimodal data such as numerical values, images, and text. However, the current power system suffers from problems such as inconsistent data formats, chaotic standards, and different interfaces, leading to difficulties in information sharing and prominent issues in multi-source data fusion. Traditional single-source data analysis methods have significant limitations in areas such as fault diagnosis, state perception, and renewable energy power prediction, making it difficult to meet the precise perception requirements of the new power system. This application's embodiment uses a cross-attention mechanism to fuse multimodal data, incorporating cross-side correlation patterns, which can accurately achieve fault diagnosis, renewable energy power prediction, and low-carbon regulation, solving the complex data processing and precise perception problems in multiple scenarios of the power system.

[0052] Figure 7This application provides an embodiment of a multi-scenario precision sensing device for power systems with high penetration of new energy sources. The device includes: The acquisition module 710 is used to acquire power data from the power source side, grid side, load side and energy storage side of the power system; The source-load linkage module 720 is used to adjust the output of the power source based on the power data of the power source side, the carbon budget constraint, and the power data of the grid side. The grid-load linkage module 730 is used to adjust the electricity consumption behavior of the load side based on the power data of the grid side and the power data of the load side if it is determined that a low-carbon demand response is triggered. The source-storage linkage module 740 is used to determine the collaborative control strategy of the power source side and the energy storage side based on the power data of the power source side and the power data of the energy storage side. The grid-storage linkage module 750 is used to optimize and control the energy storage side based on the power data of the grid side, to evaluate the response indicators of the energy storage side based on the power data of the energy storage side, and to control the energy storage side to participate in the auxiliary services of the grid side based on the evaluation results. The source-load linkage module 760 is used to determine the monthly, day-ahead, and intraday regulation of the power source and the load side based on the power data of the power source side, the carbon budget constraint, and the power data of the load side.

[0053] like Figure 8 As shown in the figure, this application provides an electronic device, including a processor 111, a communication interface 112, a memory 113, and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114. Memory 113 is used to store computer programs; In one embodiment of this application, when the processor 111 executes a program stored in the memory 113, it implements the method provided in any of the foregoing method embodiments, including: Acquire power data from the power source side, grid side, load side, and energy storage side of the power system; The power output of the power source is adjusted based on the power data from the power source side, carbon budget constraints, and power data from the power grid side. If a low-carbon demand response is triggered based on the power data from the power grid side and the power data from the load side, the electricity consumption behavior of the load side is adjusted based on the demand response incentive mechanism. Based on the power data from the power source and the power data from the energy storage side, a coordinated control strategy for the power source and the energy storage side is determined. The energy storage side is optimized and controlled based on the power data of the grid side, and the response indicators of the energy storage side are evaluated based on the power data of the energy storage side. Based on the evaluation results, the energy storage side is controlled to participate in the auxiliary services of the grid side. The monthly, day-ahead, and intraday controls for the power supply and load sides are determined based on the power data from the power source side, the carbon budget constraint, and the power data from the load side.

[0054] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method provided in any of the foregoing method embodiments.

[0055] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0056] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0057] The above embodiments are merely illustrative examples and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this application.

Claims

1. A method for multi-scenario accurate perception of a new energy high-penetration power system, characterized in that, The method comprises: acquiring power data of a power supply side, a power grid side, a load side and an energy storage side in a power system; adjusting output of the power supply side based on the power data of the power supply side, a carbon budget constraint and the power data of the power grid side; if it is judged that a low-carbon demand response is triggered based on the power data of the power grid side and the power data of the load side, adjusting electricity consumption behavior of the load side based on a demand response incentive mechanism; determining a coordinated control strategy of the power supply side and the energy storage side based on the power data of the power supply side and the power data of the energy storage side; optimizing and controlling the energy storage side based on the power data of the power grid side, and evaluating a response index of the energy storage side based on the power data of the energy storage side, and controlling the energy storage side to participate in auxiliary services of the power grid side based on an evaluation result; determining monthly regulation, day-ahead regulation and intra-day regulation of the power supply side and the load side based on the power data of the power supply side, the carbon budget constraint and the power data of the load side.

2. The method of claim 1, wherein, The power data is multi-modal data, and the multi-modal data comprises text modal data, frequency domain modal data, image modal data and numerical modal data; The method further comprises: extracting features from the multi-modal data, and performing feature processing to obtain corresponding text modal features, frequency domain modal features, image modal features and numerical modal features; for any one of the power supply side, the power grid side, the load side and the energy storage side, determining same-modal cross-side interaction features based on the text modal features, the frequency domain modal features, the image modal features and the numerical modal features by using a cross-attention mechanism, and determining corresponding single-modal fusion features based on the text modal features, the frequency domain modal features, the image modal features, the numerical modal features and the corresponding same-modal cross-side interaction features, and determining corresponding global fusion features based on the single-modal fusion features and corresponding weights; splicing the global fusion features corresponding to the power supply side, the power grid side, the load side and the energy storage side respectively to obtain spliced features, and performing fault detection, new energy power prediction or low-carbon regulation based on the spliced features.

3. The method of claim 2, wherein, Determining same-modal cross-side interaction features based on the text modal features, the frequency domain modal features, the image modal features and the numerical modal features by using a cross-attention mechanism comprises: sequentially taking features of one modality in the text modal features, the frequency domain modal features, the image modal features and the numerical modal features as target modal features; taking the target modal features as a query vector of the cross-attention mechanism, taking a feature vector obtained by column splicing of remaining modal features as a key vector of the cross-attention mechanism, and taking a feature vector obtained by row splicing of the remaining modal features as a value vector of the cross-attention mechanism to obtain same-modal cross-side interaction features corresponding to the target modal features.

4. The method of claim 1, wherein, The method for adjusting output of the power supply side based on the power data of the power supply side, a carbon budget constraint and the power data of the power grid side comprises: determine new energy output, line flow redundancy and carbon budget surplus based on the power data of the power supply side and the power data of the power grid side; in the case that the change amplitude of the new energy output and the new energy output at the last time is greater than the preset amplitude threshold, and the line flow redundancy is less than the preset redundancy threshold, dispatch the reactive power compensation device of the power supply side, adjust the output of the hydropower station of the power supply side, and control the charging of the energy storage side; if the new energy output is lower than the preset output threshold and the carbon budget surplus is less than the preset target threshold, control the low-carbon operation of the thermal power unit of the power supply side and optimize the line flow transmission path of the power grid side.

5. The method of claim 1, wherein, if it is judged to trigger low-carbon demand response based on the power data of the power grid side and the power data of the load side, adjust the power consumption behavior of the load side based on the demand response incentive mechanism, including: determine the power grid side load rate and the proportion of low-carbon power output based on the power data of the power grid side and the power data of the load side; if the power grid side load rate or the proportion of low-carbon power output meets the preset condition, trigger low-carbon demand response, push peak load reduction instructions and low-valley production transfer suggestions to industrial users, push low-carbon power consumption time and low-valley charging incentives to residential users, increase the compensation standard for the residential users under the condition that the load adjustment amount of the power supply side does not meet the preset adjustment amount, send the demand response income bill to the residential users, and classify the residential users based on the response rate of the residential users.

6. The method of claim 1, wherein, determine the coordinated control strategy of the power supply side and the energy storage side based on the power data of the power supply side and the power data of the energy storage side, including: if the predicted value of new energy output fluctuation is greater than the prediction threshold, control the charging and discharging of the energy storage side; if the battery health of the energy storage side is less than the preset health threshold, control the charging and discharging depth of the energy storage side; if it is judged that the energy storage side has overcurrent or overtemperature based on the power data of the energy storage side, switch the standby energy storage unit of the energy storage side; if the fluctuation rate of the linkage of the power supply side and the energy storage side does not reach the preset standard, optimize the charging and discharging power distribution of the energy storage side.

7. The method of claim 1, wherein, optimize and control the energy storage side based on the power data of the power grid side, including: if it is judged that the line flow value of the power grid side is greater than the rated capacity ratio, control the energy storage side to charge; if the voltage deviation of the transformer substation is greater than the qualified threshold, control the energy storage side to provide reactive power support, and if the power supply side fails, control the energy storage side to discharge; based on the line flow distribution of the power grid side, optimize the deployment location and standby capacity of the energy storage side.

8. The method of claim 1, wherein, determine the monthly regulation, day-ahead regulation and intra-day regulation of the power supply side and the load side based on the power data of the power supply side, the carbon budget constraint and the power data of the load side, including: determining a consumption rate of a monthly carbon budget based on the power data of the power supply side and the power data of the load side, determining a target of monthly low-carbon power output of the power supply side based on the monthly carbon budget and an electric carbon factor of the power supply side if the consumption rate exceeds a pre-designed plan value, controlling a monthly load capacity capable of participating in peak shaving to match the target of the monthly low-carbon power output of the power supply side, integrating industrial load and residential load into a load aggregation based on an adjustable load potential of the load side, and sending a monthly regulation benchmark to the power supply side and the load side; determining a day-ahead collaborative plan of the power supply side and the load side based on a new energy output prediction value of the next day and a change of an electric exploration factor, pushing a load reservation instruction of the next day to the industrial load, and pushing a low-carbon power consumption period of the next day to the residential load, and adjusting an output plan of the power supply side based on a load amount capable of participating in adjustment of the next day; if new energy output growth leads to local load deficiency, sending power consumption incentives to the load aggregation to guide consumption of excess new energy output, and if grid side congestion occurs, controlling a micro-grid in a congestion area to switch to an island operation mode to preferentially consume local new energy output.

9. A multi-scene accurate perception device for a new energy high-penetration power system, characterized in that, The method comprises the following steps: acquiring power data of a power supply side, a grid side, a load side, and an energy storage side in a power system; adjusting output of the power supply side based on the power data of the power supply side, a carbon budget constraint, and the power data of the grid side; if low-carbon demand response is triggered based on the power data of the grid side and the power data of the load side, adjusting power consumption behavior of the load side based on a demand response incentive mechanism; determining a collaborative control strategy of the power supply side and the energy storage side based on the power data of the power supply side and the power data of the energy storage side; optimizing and controlling the energy storage side based on the power data of the grid side, and evaluating a response index of the energy storage side based on the power data of the energy storage side, and controlling the energy storage side to participate in auxiliary services of the grid side based on an evaluation result; determining monthly regulation, day-ahead regulation, and intra-day regulation of the power supply side and the load side based on the power data of the power supply side, the carbon budget constraint, and the power data of the load side.

10. An electronic device, comprising: The method comprises a memory and a processor, the memory stores a computer program, and the processor executes the computer program to implement the method in any one of claims 1-8.

Images